|
🧰 Google ADK: Agent Development Kit
Last week, we learned why Google waiting until many other agent orchestration frameworks might have given them a leg up in the core constructs they used in the Agents SDK. Spoiler alert: it's pretty good!
|
|
|
|
|
|
Building Agentic Systems with Protocols: MCP & Agent2Agents (A2A)
Tomorrow at 2:00 PM PT, we're diving deep into protocols including both MCP and A2A! How should we think about building production agentic applications with these protocols? When do we need them, exactly? How are they useful to our end users, customers, or internal stakeholders? Also, we'll build our own MCP server!
|
|
|
A2A: Agent2Agent Protocol
On Wednesday, June 4, at 10:00 AM PT, we'll go even deeper into Agent2Agent (A2A) Protocol to describe exactly what it is, how it's defined (according to the spec), and how we should think about leveraging it as we align our production LLM applications with the emerging standards of the industry
|
|
|
Join the AI Engineering Bootcamp Cohort 6 Demo Day!
On Thursday, June 5 at 4 PM PT, join us live for Demo Day to see what students have built, shipped, and are ready to share 🏗️ 🚢 🚀! Check out projects here.
The aspiring AI Engineers in Cohort 6 have been working hard for the past 10 weeks building, shipping, and sharing production LLM applications at the open-source edge. Demo Day is where they'll showcase their work. See you there!
|
|
|
Codex: The Autonomous SWE
On Wednesday, June 11, at 10:00 AM PT, join to tour ChatGPT Codex, OpenAI’s first cloud-hosted Autonomous Software Engineer that reads, edits, tests, and commits code while you sleep—going far beyond the lightweight Codex CLI agent we met last month. In this session, we’ll unpack its sandboxed terminal architecture, see how technical teams already launch dozens of parallel Codex instances per hour, and learn best-practice playbooks for folding A-SWEs into real-world PR pipelines.
|
|
|
RAG: The 2025 Best-Practice Stack, Prototype to Production
Ready to turn RAG from “table-stakes prototype” into a production powerhouse? On Wednesday, June 18, at 9:30 AM ET, Dr. Greg and The Wiz will be at the Toronto Machine Learning Summit where they will first reveal the 2025 best-practice stack—LangGraph + Qdrant + Cohere Rerank + RAGAS—then show how to harden that build, win stakeholder buy-in, and scale it from on-prem demo to cloud or air-gapped deployment without rewriting a line of code. Plus, they'll do a live session from TMLS - stay tuned for more details!
|
|
|
From Text to Vectors: Mastering Tokenization and Embeddings for Transformer-Based AI Systems
Tokens aren’t just syntax — they’re the first step in transforming raw language into high-dimensional vectors that fire up every Transformer you deploy. Saumil Srivastava's latest post walks you through that journey and packs in battle-tested tips engineering leaders can use right away to balance performance, cost, and accuracy when building AI systems.
|
|
|
Multi-Agent AI: When to Choose LangGraph vs OpenAI Assistants
Struggling to choose between the state-rich, graph-orchestrated power of LangGraph and the lightweight, thread-based convenience of OpenAI Assistants? Suhas Prakash's latest post distills the trade-offs—control, cost, tooling, and scale—into a quick-read guide so you can pick the perfect backbone for your next multi-agent build.
|
|
|
LangGraph Deep Research: A tale of two architectures
Enterprise-grade multi-agent workflows or a lean, privacy-first loop—Don Branson's latest post dissects LangGraph’s Open Deep Research and Ollama Deep Research builds so you know when to choose each. Dive in to see how graph-orchestrated scale stacks up against local simplicity for your next AI project.
|
|
|
|
|
|
The AI Engineering Bootcamp, Cohort 7 |
|
|
Cohort 7 of The AI Engineering Bootcamp kicks off June 24th with practitioners around the world. Transform your skillset and become an AI Engineer in 2025.
|
|
|